基于深度学习的语音情感识别方法研究  

Research on Speech Emotion Recognition Method Based on Deep Learning

作　　者：郭晓琳 GUO Xiaoin(Yunnan Economics Trade and Foreign Affairs College,Kunming 650000,China)

机构地区：[1]云南经贸外事职业学院,云南昆明650000

出　　处：《电声技术》2024年第4期48-50,共3页Audio Engineering

摘　　要：为提高情感识别任务的准确性和效率,探讨了基于深度学习的语音情感识别方法。首先,构建一个综合的深度学习框架,用于处理语音情感识别任务。其次,深入研究一维卷积神经网络(Convolutional Neural Networks,CNN)在语音情感分析中的应用,探讨其在特征提取和情感分类中的优势。最后,进行实验分析。实验结果表明,该方法能够有效识别语音情感,具有一定的稳定性和可靠性。To improve the accuracy and efficiency of emotion recognition tasks,this article explores speech emotion recognition methods based on deep learning.Firstly,construct a comprehensive deep learning framework for processing speech emotion recognition tasks.Secondly,conduct in-depth research on the application of one-dimensional Convolutional Neural Networks(CNN)in speech sentiment analysis,and explore its advantages in feature extraction and sentiment classification.Finally,conduct experimental analysis.The experimental results show that this method can effectively recognize speech emotions and has a certain degree of stability and reliability.

关 键 词：深度学习 卷积神经网络(CNN) 语音分析 情感识别 

分 类 号：TN912.34[电子电信—通信与信息系统]